Artificial Intelligence for Law Enforcement and Community Safety (AiLECS) Lab
Chief Investigators: Dr Campbell Wilson, Janis Dalins
In collaboration with the Australian Federal Police (AFP), this Australian-first Artificial Intelligence for Law Enforcement of Community Safety (AiLECS) Lab builds on prior collaboration around research projects on automated classification of distressing materials, characterisation of behaviour on the dark web and prioritised file search. It has an overarching theme of the ethical use of artificial intelligence in law enforcement.
This Lab will investigate how AI and machine learning can assist with accelerating the investigation and prosecution of heinous crimes such as child exploitation and terrorism, exposure to which also causes significant psychological trauma for police officers and the judiciary.
The Lab will have an initial focus on:
- Explainable algorithms and auditable performance techniques (important for judicial use and public accountability)
- Illicit image and video machine learning classification algorithms
- Image localisation
- Scalable near duplicate image detection techniques
- Techniques for leveraging massive law enforcement datasets.
A crucial part of this work has been access to real case data for training machine learning algorithms. Due to its illegal and disturbing nature, a “data airlock” server architecture has been developed (by Data61, the AFP and Monash), allowing tightly controlled access in such a way that researchers can train algorithms without being exposed to the raw data. The airlock will be used by researchers worldwide, particularly those working with law enforcement and possibly large platform providers where content moderation is an issue.
A collaboration between: